Industrial software company TwinThread has launched an AI-powered analytics solution designed to help manufacturers improve batch consistency, reduce production cycle times and identify unused capacity in existing equipment.
The Perfect Batch platform analyses historical production data to establish optimal batch profiles, which the company calls “golden batches”. It then recommends adjustments to help manufacturers replicate better-performing production runs across different recipes, shifts and raw material variations.
The launch reflects growing interest in applying industrial AI to improve manufacturing performance without requiring investment in additional production equipment. However, the potential gains will depend on the quality of historical data and manufacturers’ ability to implement the system’s recommendations.
Perfect Batch connects to existing batch execution systems and uses historical data to build digital twins and train analytical models. TwinThread says the initial modelling can be completed in hours, allowing manufacturers to begin identifying opportunities for improvement more quickly than with conventional approaches.
Rather than relying on manually configured control limits, the software learns process parameters from historical performance. It monitors five areas: equipment utilisation, cycle time, quality, process control and yield.
The platform also analyses individual production cycles to identify bottlenecks and periods of lost production time. Automated alerts and diagnostic tools are intended to help operators investigate exceptions rather than spend time manually reviewing routine production data.
For manufacturers seeking greater automation, TwinThread offers an optional integration with Thread Builder, its real-time workflow engine. This can automate anomaly responses, carry out diagnoses and trigger corrective actions, although the extent of intervention will depend on how the system is configured.
The company is also targeting regulated industries, where the software can track materials, monitor quality and yield against requirements, and maintain records for audits.
“[The] core promise is to empower operations teams to deliver the ideal batch every time, regardless of recipe, shift, or variations in raw materials,” said Erik Udstuen, Co-Founder and Chief Executive of TwinThread.
He said the system was designed to uncover unused production capacity and help manufacturers improve the financial performance of existing assets.
Beyond individual production lines, Perfect Batch provides a consolidated view of equipment utilisation and batch performance across multiple manufacturing sites. TwinThread says this could help operations, engineering and supply chain teams compare performance and prioritise improvement projects across their production networks.
The software is available in Simple Batch and Complex Batch versions, aimed at different levels of production complexity.
TwinThread describes Perfect Batch as part of its wider industrial AI platform, which combines predictive, prescriptive, generative and agentic AI technologies. The company says its platform has more than one million digital twins, nearly 50 million sensors and two million AI models deployed.
